Developing a Data-Driven Tiered Instructional Model for Advanced Mathematics Learning in Engineering Education: A Case Study in a Chinese University

Despite widespread adoption of tiered instruction in engineering mathematics education, existing approaches lack a systematic integration of learning analytics and data-driven decision-making frameworks; instead, they rely on subjective teacher judgments and static ability groupings that fail to address the dynamic, multidimensional nature of student heterogeneity. This study developed and evaluated a comprehensive data-driven tiered instructional model specifically designed for advanced mathematics learning in engineering contexts. The intervention included four core components- (a) multidimensional diagnostic assessment capturing students' prior knowledge, mathematical thinking skills, and motivational profiles; (b) evidence-based stratification using priority needs index analysis and thematic coding of qualitative data; (c) adaptive three-tier instructional delivery (Basic, General, Development) with modular resources tailored to different proficiency levels; and (d) continuous progress monitoring with real-time instructional adjustments based on formative assessment data. A quasi-experimental study with first-year engineering students revealed significant benefits of the data-driven model. Students receiving tiered instruction substantially outperformed their conventionally taught peers on mathematics achievement (mean difference = 10.67 points, 95% CI [2.48, 18.85], p = 0.012, Cohen's d = 0.67), indicating a medium-to-large effect size. The model effectively reduces achievement gaps while promoting excellence. This study contributes a validated framework for implementing data-driven tiered instruction in undergraduate mathematics education, demonstrating that the systematic integration of these strategies can enhance learning outcomes across diverse student populations.